Terms of Trade and Economic Growth in a World of Constrained Capital Mobility

Terms of Trade and Economic Growth in a World of Constrained Capital Mobility

Hildegunn Kyvik Nordås

WP 1997: 5

.

Inl

W orking Paper

Chr. Michelsen Institute

Development Studies and Human Rights

Bergen Norway

Terms of Trade and Economic Growth in a World of Constrained Capital Mobility

Hildegunn Kyvik Nordås

WP 1997: 5 Bergen, March 1997

l-i ~e~~~~~~~~:~~~ ~~:2~~~E

W orking Paper WP 1997: 5

Terms of Trade and Economic Growth in a World of Constrained Capita) Mobility

Hildegunn Kyvik Nordås

Bergen, March 1997

Summary:

This paper focuses on the interaction between world community and capital markets within the

framework of an extended neoclassical growth mode!. The model incorporates raw materials as an essential input and captures the observed, but hitherto unexplained impact of terms of trade on economic growth. It is shown that in a world of declining real prices of industrial raw materials, the

steady state growth rate is an increasing function of the input ratio raw materials/unskilled labour.

When international capital flows are constrained, the speed of convergence is determined by the same variables.

Sammendrag:

Interaksjonen mellom internasjonale råvare markeder og den internasjonale kapitalmarknaden blir

analysert innanfor ramma av nyklassisk vekstmodell. I modellen inngår råvarer som essensielle innsatsvarer i produksjonen. Modellen viser korleis bytteforholdet påverkar den økonomiske veksten. Råvareimportørar som opplever forbetring av bytteforholdet vil ha høgare vekst i "steady state" jo større del av innsatsfaktorane råvarer utgjer relativt til ukvalifisert arbeidskraft. Dersom

den internasjonale kapital marknaden er imperfekt, vil også konvergeringsraten avhenga positivt av

forholdet råvarer/ukvalifisert arbeidskraft.

Indexing terms:

Economic growth Capital mobility Natural resources Terms of trade JEL 03, 04, 013, 016

Stikkord:

Økonomisk vekst Kapitalmobiltet Råvarer

To be orderedfrom Chr. Michelsen Institute, Fantoftvegen 38, N-5036 Fantoft, Bergen, Norway.

Telephone: +4755574000. Telefax: +4755574166

Contents

1. Introduction ¹

2. Relations to previous research 3

3. The Ramsey model

3.1 The Ramsey model with no international capital flows 3.2 The model with international capital flows

3.2.1 Transitionardynamics

5 8 12 15

4. Some empirical evidence

4.1 Growth and resource intensity

4.2 Foreign direct investment and local investment in

immobile capital 5. Summar and conclusions

18 18

21 23

1. Introduction

Empirical work on the determnants of growth has found that changes in terms of trade have a significantly positive impact on economic growth (Barro and Sala-i- Marin 1995). This relationship has, however, not been captured by theoretical growth models. To the contrar, as observed by Barro (1996), changes in terms of trade should not affect real GDP growth unless it stimulates employment. This paper develops a model which entails a channel through which terms of trade affect

economic growth. A precondition for the presence of a relationship between terms of trade and growth is the existence of a long-run trend in terms of trade. Such a trend is

most clearly seen between non-oil commodity-exporting countries and the

industrialized countries since the 1960s as shown in figure 1.1

Figure 1

Price index of metals and minerals, constant 1990 US dollars 200

50 150 100

o

~~~~~ro~ n~N~~n~rn~~ ~~M~$~~~OO~ ~~~

Source: World Bank 1994

This paper focuses on the impact of this trend on countries who use imported raw materials as inputs in industrial production and export manufactured goods. It extends

i There has been some controversy over the development in real commodity prices. The Hotelling rule states that the real price of exhaustible natural resources wil increase at a rate equal to the real interest rate over time, while the Prebish - Singer hypothesis asserts that commodity prices are on a long-term

the neoclassical growth model by the inclusion of industrial raw materials as essential inputs in the productionJunction. Earlier work on the oil price shocks in the 1970s and 1980s emphasized the role that international capital markets played in absorbing the shocks and redistributing world income (Bruno and Sachs 1985, Christini 1995).

Therefore, in order to capture the long-run relationship between world commodity and capital markets, the produetion function also entails two classes of capital

distinguished by the degree of international mobilty. This setting provides a

framework for analyzing how world commodity- and capital markets interact with local investment to generate economic growth, and engenders additional insight as to

why growth rates differ among countries. It finally contributes to a better

understanding of why capital does not flow to labor-abundant developing countries to the extent predicted by the standard neoclassical growth mode!.

I show that the declining price trend depicted in figure 1 has the same impact on economic growth as has labor-augmenting technological progress. Further, it is shown that the steady state growth rate is determined by the relative importance of raw

materials as compared to raw labor in the production process. The larger the ratio of raw materials to raw labor, the higher the growth rate in a situation with declining raw materials prices. This ratio varies among countries and is a function of industrial structure and technology.

The interaction between world capital and commodity markets works as follows:

Declining real costs of raw materials improve the productivity of all other factors of production, and more so the higher the raw materials to raw labor ratio. Therefore,

capital would flow from less raw materials intensive countries to more raw materials intensive countries in a world ofpedectly mobile capita!. With constrained capital mobility, capital flows are restricted by local investment in immobile capital. The model predicts that the fastest-growing countries in the world are relatively raw materials intensive, open to international capital flows and have a high investment rate

in immobile capital such as human capital and various forms of infrastructure. It also

helps to be a net importer of raw materials, although the net gains from declining

prices of raw materials may be positive for net commodity exporting countries also, if they have a relatively resource-intensive industrial structure and declining relative prices reflects declining relative costs in the extractive and basic processing industries.

The rest of the paper is organized as follows: Section 2 briefly review related

research, while section 3 presents the Ramsey growth mode!. The simplest version for an economy closed to international capital flows is discussed in section 3.1 and an extension with constrained international capital flows is laid out in section 3.2. Some empirical evidence is presented in section 4, and section 5 concludes.

2. Relations'to previous research

The interrelationship between commodity prices and economic activity has previously been analyzed by Bruno and Sachs (1985) and Christini (1995). Both contributions

focused on the impact of a commodity price shock on world real output and

employment and both emphasized the role of world capital markets as a transmitter

and absorber of the price shock. Bruno and Sachs (1985) allowed commodity prices

to influence output through two channels.First, changes in commodity prices shiftthe factor price frontier in a similar way as Hicks-neutral technological change. Second, commodity prices affect demand through real wages,employment and the world distribution of wealth. Christini (1995) focused on demand effects only. I incorporate the insights from Bruno and Sachs (1985) into a neoc1assical growth framework and analyze how industrial structure, represented by the intensity of raw materials use, affects steady state growth as well as the transitionar dynamics.

Raw materials and energy are routinely incorporated into macro-economic forecasting and business cycle models, but are absent from growth models. There is, however a growing literature on the role of intermediate inputs in growth and development,

focusing either on productivity gains from increased specialization through an

expanding variety of inputs (Romer L 987) or improvements of the quality of inputs (Grossman and Helpman 1991 ch. 4). In these models intermediate inputs are the

vehicle for technological progress. Raw materials differ from such intermediate

inputs in three important ways. First, they are homogeneous and standardized and are not seen as cariers of new technology. Second, they are internationally tradable,

while access to a limited range of specialized intermediate products has often been

stated as a cause of poverty traps in growth models.2 Third, they are to a large extent provided by nature. Therefore, raw materials have more in common with raw or unskilled labor than differentiated intermediate inputs, and are accordingly lumped together with labor into a composite non-accumulated factor of production in the

2 Recent contributions are Rodrik (1996), and Rodriguez-Clare (1996).

model developed in this paper. In this regard the model can be seen as a first step

towards integrating business cyc1e and growth models as far as treatment of raw

materials is concerned.

A growth model with constrained capital mobility was developed by Baro, Mankiw and Sala-i-Marin (1995). The innovation of this paper is to combine their model with a composite non-accumulated factor of production consisting of unskilled labor and industrial raw materials into a two level Cobb Douglas production function. Within this framework the dynamic interaction between international commodity- and capital markets in a world of constrained capital mobilty is analyzed.

3. The Ramsey model

The setting is a Ramsey growth model for a small, open economy. The production function is a two-level nested Cobb-Douglas function, as the special case of a two- level nested CES function. At the first level the production function contains two inputs, non-accumulated factors, denoted N, and accumulated factors, denoted D. At

the sec ond leve!, non-accumulated factors are split into labor and raw materials, while

accumulated factors are split into mobile and immobile capita!. Consequently, each

composite consists of one internationally mobile and one immobile factor of

production as raw labor is considered internationally immobile. The production function is given by:

y= ANI-uDa (1)

Where y is gross output and A is the level of technology. In intensive form the production function reads:3

y = Ad a (1')

where y is output per unit of non-accumulated factors and d is the ratio of

A

accumulated to non-accumulated factors of production. Income generated can be

spent on consumer goods or saved. The allocation between consumption and savings

is determined by the consumers' utility maximization problem. Infinitely lived households maximize the standard constant intertemporal elasticity of substitution

utilty function:

- CH1 -1

u(c) = i e-P1dt

I i-e ⁽²⁾

where c is consumption per capita, p is the time preference rate and e is the elasticity

of marginal utility. It is assumed that households hold assets in the form of claims on capita!. In addition each person supplies inelastically one unit of labor services per

unit of time, while a quantity of raw materials per unit of time corresponding to

demand at the exogenously determined average price during that unit of time is imported. For convenience it is assumed that the population and the labor force remain constant. The equilibrium capital accumulation trajectory together with the

3 Lower case letters refer to variables in per capita units while lower case letters with a hat refer to variables measured in units of non-accumulated factors throughout the paper.

equilibrium growth path of income and consumption are found through the usu al

procedure where consumers maximize utilityand firms maximize profits, subject to a set of constraints. Let us star with the consumer. She maximizes utility as given by equation (2) subject to the flow household budget constraint:

å = ra + w - c ₍₃₎

where a is the stock of assets per capita, r is the rate of return on savings, w is the

wage rate. The utility maximization problem can be sol ved by means of the present

value Hamiltonian:

C i-e -1

J = e -pr + v(ra + w - c)

i-e

⁽⁴⁾

From the first-order conditions we get the optimal consumption path:

ë 1

- = - (r - p)

c e

⁽⁵⁾

which is the standard Euler equation of Ramsey growth models. The equilibrium rate of interest depends on the structure of the capital market, and is derived for each case in the subsections below.

3.1 The Ramsey model with no international capita

i flows

In thissubsection the aggregateaccumulated capital input, D, is treated as one

homogeneous input, while the non-accumulated composite of labor and raw materials is given by:

N = L1-ßQß (6)

where L is labor and Q is raw materials. Plugging this into the production function (l) yields:

y= A(L1-ßQßy-a Da (1 a)

In intensive form (1') still applies. Firms maxiffze present value profits, whicn

reduces to maximizing profits in each period of time in the absence of adjustment

costs. Profits are given by:

1C=A(Li-ßQß)I~aDa -(rd +Ô)D-wL-PqQ (7)

where Ô is the depreciation rate of accumulated factors and rd is the net renta! rate on accumulated factors and P q is the exogenously given price of natural resources measured in units of final output. Before proceeding with finding the equilbrium trajectory of this economy one important property of the model is observed:

Observation 1

Given the production function(l a),decliningcommodity prices have the same impact

on aggregate output as has labor-augmenting technological progress.

To see this, note that demand for raw materials can be deri ved by maximizing profits

with respect to Q. This yields the familiar demand schedule

Q = ß(l-a)Y / Pq

(8)

Substituting back into (1 a) gives:

y = Ã( L1-ß pq-ß r-a)ç DaÇ (lb)

- ( ß) J1/0-ß(I-a))

where A= A(ß(1-a) (I-a and ç=lI(1-ß(1-a)). Assume that Pq

dec1ines over time. Then (1 b) has the form Y(t) = F(A(t)L, K(t)), which is the general production function with labor-augmenting technological progress.

In order to find the equilibrium growth path of the economy, it is useful to express the system in terms of variables that are constant in steady state. Observation L implies

that non-accumulated factors have the same properties in the model as has "efficient labor" in standard neoc1assical growth models. Therefore, the variables in terms of non-accumulated factor units are constant in steady state. With no international capital flows a = d in equilibrium. Further, in equilibrium the rental rates or prices of inputs must equal their marginal products. Now, let Pq change at a constant,exogenous rate

ry and recall that the labor force and population are assumed to be constant. From (8) . wehave thatQ, andhence the ratio Q/L, grow at the rate ~ry while Ngrows at the rate

-ßry. Then the profit equation (7) can be rearanged to read:

7C= N(Ada -(rd +D)d-w(Q/ L)ß -Pq(Q/ L)I-ß)

_(T)

At time t, the ratio Q / L = (Qo / Lo)e -!Jr, where we set (Qo / Lo) = ß / (1- ß). The

first order conditions for profit maximizing read: rd = aAda-1 -D, w=(1-a)(1-ß)Adae-ß!J1 and Pqq=ß(l-a)Adae-ßryr. Substituting all this into

the flow household budget constraint yields the overall economy resource constraint,

which in turn determines the accumulation of capital per unit of non-accumulated

factors:

d = (l - (1- a)ß)Ada - ê - (D - ßry)d (9)

where we have subtracted the share of gross output paid to foreign suppliers of raw materials. Finally, inserting the equilbrium rental rate of capital into the Euler

equation and using the condition that ê =C / N = ceßry, we have that

ê c 1 ( Aa i )

-; = - + ßry = - (1- (1- a)ß)aAd - - D - p + 8ßry

c c 8

⁽¹⁰⁾

In steady state y / Y = ¿ / ê = a(d. / d.) = o. Therefore, gross output and the stock of

accumulated factors, both in per capita terms, grow at the rate -ß17 and -ß17/a

respectively, which are positive constants as long as 17 .c O. Jf prices of raw materials

should follow the Hotellng rule and increase at a rate r, on the other hand, we have a

"limits to growth" scenario, where raw materials become increasingly scarce, and consequently output and consumption dec1ine. These results are summarized as follows:

Result 1

In steady state

i) l =!: = -ß17 ). O whenever 17 .c O, ii) d = - ß17 ). O whenever 17 .c O

y c d a

Note that the only thing that matters for steady state growth in this setting is the composition of the non-accumulated production factor, that is the relative importance

of raw materials compared to unskiled labor. Jf we allow the value of ß to var

among countries, growth rates wil differ too. An interpretation of this result is that countries grow faster in steady state the higher the share of raw materials intensive

industries compared to (unskilled) labor-intensive industries in total output. It also

suggests that countries grow faster the larger the share of goods-producing sectors in

total output.

3.2 The model with international capital flows

Twomodifications to the modelare introduced in this section. First, theaccumulated factor, D, is a composite of mobile capital denoted K, and immobile capital denoted H. Second, we allow for international capital flows, but only in mobile capital as the designation suggests. Internationally mobile capital earns the same rate of return in all countries, which is taken to be exogenous. The return to H-capital, on the other hand, may var among countries. However, since investors always have the opportunity to

invest abroad, the expected return to new investments in H-capital can not be lower than that on K-capital. The capital composite is defined as:

D = Hi-r Kr (11)

Plugging this into the production function (1) yields

y = AN l-a

(HI-r Krr

^(lb)

The stock of K-capital employed in any economy is determined by the condition that its marginal product equals the world interest rate at every point in time. K-capital employed is thus not restricted by domestic savings, and is given by:

K= ra y

r+Ô

⁽¹²⁾

where r is the exogenous international rate of interest. Clearly, the K-capital output

ratio is constant over time whether the economy is in steady state or not..Plugging (12)

back into the production function (1b) yields Y = AN (l-al/(Hryl Ha(l-yi/(l-ari, or in

intensive form:

y = AhE (1b')

where A is a constant defined by A = A(ar / r)ar :;

- ( )iii-ar

A, and

ê = a(1- r) / (1- ar).o a Note that output is now a function of the ratio of H-

capital to non-accumulated factors only. To determne accumulation of H-capital, we again turn to the consumer and her utilty maximization problem. She now maximizes (2) subject to the flow household budget constraint broken down as follows:

ii + ak. + åk = rh h + rak + w - c (13)

where a is the share of K-capital owned by domestic citizens. Whenever rh :; r, all domestic savings wil be invested in H-capital, and cr = O. In that case the constraint on foreign investment is binding. When rh = r on the other hand, domestic savings wil be allocated among the two kinds of capital such that h/ = (l-y)/y, or invested abroad. In the following it wil be assumed that initially rh :; r and hence cr = O in

developing countries, while rh = r in developed countries. It follows that the

demarcation line between developed and developing countries is whether return to H-

ratio h/ is lower than or equal to the ratio (l-y)/y. The resource-constraint for the

representativehousehold in adeveloping country then reads:

h=rhh+w-c

^(13a)

The resource constraint for the economy as a whole in terms of units of non-

accumulated factors is found by substituting the equilbrium factor prices into (13a) and maintaining the result from the previous section that the Naggregate grows at a rate -ßTJ :: O due to dec1ining raw materials prices:

h = (1- (1- a)ß)(l- ar)Ah£ - ê - (o - ßTJ)h (13b)

For convenience it is assumed that the depreciation rate is the same for H- and K-

capita!. Steady state consumption per unit of non-accumulated factors is now given

by:

ê c l ( _A£ i )

-; = - + ßTJ = - (1- (1- a)ß)(1- ar)êAh - - o - p + ()ßTJ

c c ()

⁽¹⁴⁾

Again we have a steady state characterized by 2/ ê = Y / jì = e(h / h) = O, while per

capita output grows at a rate -ßTJ. Note that since there is a non-accumulated factor of production, this is not a model of endogenous growth in spite of the constant h/ ratio.

Comparing (14) to (10) it is clear that the steady state growth trajectory is unaffected

by whether capital is internationally mobile or not. What is interesting to developing countries, therefore,are thetransitionary dynamcs, to which we now turn.

3.2.1 Transitionary dynamics

Given that K is perfectly mobile and K and H are perfect substitutes as a store of value, we have that rh = r in steady state. The transition path is characterized by an increasing hi ratio from an initial leve1 below (1-y)/y, meaning that local investment in H-capital poses a constraint on net inflows of K-capital. As commodity prices

dec1ine, the marginal product of all other inputs increases and induces an

instantaneous inflow of investment to restore equilbrium in the international capital market. In addition, as the marginal return to H-capital increases, local investors add

to their stock as long as rh :; r, triggering a sec ond wave of capital inflows.

The steady state leve1s of output per unit of non-accumulated factors are deri ved from

equations (10) and (1') for the c10sed economy and (14) and (1b') for the open economy:

A* = A JI(~a) ( / (ô _ eR ))af(l~a)

y closed a + P iA1

^(15a)

Y:pen = A if(l~e)((1_ aY)ê / (Ô + p - eß17) t(l-) (15b)

These conditions imply that the steady state level of income per unit of non-

accumulated factors, and consequently income per capita islower the larger the share

of raw materials in the non-accumulated factor composite as lon g as 17 .c O. Note also

that for reasonable parameter values Y;/osed :; Y;pen' To keep the analysis tractable, let

us make the simplifying assumption that the savings rate remains constant during the

transition period. Then, as show n by Baro and Sala-i-Marin (1995) the speed of

convergence is given by:4

Vclosed = (1- a)(c5 - ßr¡) (16)

V open = (1- e)(c5 - ßr¡) (17)

As is well known, the open economy converges towards steady state faster than the c10sed economy, and consequently it converges faster the larger the share of K-capital

in total accumulated factors, which in turn is reflected in a smaller e. Further, the speed of convergence is faster the less capital intensive and the more raw materials intensive the technology, again provided that r¡ oc O. Note that a and ß are independent, such that an economy may be both raw materials intensive and capital~

intensive, which in fact is often the case.5 In other words, by our definition of a developed country, we have that a developing country becomes developed in a shorter period of time the more open it is towards foreign investment, and the more resource- intensive its technology. It does, however, become "developed" at a lower leve! of income. These results are summarized in result 2 and ilustrated by figure 2.

4 The restrictions imposed on parameters for a constant savings rate is fl = (c5 + p) / (ec5 + ßr¡(1- e))

5 See Nordås (1996) for a case study of the South African economy.

Figure 2

7 T

ca

6 I

-

^{5 i}

eica

Co.. 4

elc- a. ₃

c "

c 2

1

o

Result 2

Growth trajectory for the open economy

.,...,...

"...,...,...

...,,

",

",

"

"

..I"

high-beta case . . . - - . low-beta case

time

When 17 c: O, jì is lower and v faster the larger is ß.

Figure 2 represents two developing countries staring out at the same initial leve! of output per capita and they are similar in all respects except for the value of ß. The

graphs are drawn from a linearized convergence trajectory where

ln(y(t)) = e-VI In y(O) + (1- e-VI) ln(y*) , and y = y(ß / (1- ß))ß e-ryr. The high-ß

country embarks on a rapid convergence path, and reaches a steady state while the

low-ß country is stil in its transitionar phase. For the parameter values applied in

the figure, the steady state growth rate for the high-ß country is above the growth rates during the transitionar period in the low-ß country, such that the real income gap widens during the entire period.6 For a trend dec1ine of real commodity prices of 2

percent annually, the transitionarperiod takes about 200 years for the high-ß country,

underscoring theimportance of foclising on transitionar dynamics.

4. Some empirical evidence

4.1 Growth and resource intensity

The results in section 3 imply that real GDP growth depends on terms of trade. In this section the interrelationship between real prices of imported raw materials and real GDP is investigated empirically. Table 1 ilustrates the importance of the cost of raw materials as a source of growth in some selected countries during the period 1970-92.

Annual average changes in real Pq are taken to be the annual average change in Pq measured in local currency deflated by the local GDP deflator.7 Calculating ß is more problematic. While Q is taken to be the nominal value of imported metals and minerals (SITC code 27, 28, and 68) in local currency, it is more difficult to ca1culate the value of unskilled labor input since national accounts statistics do not inc1ude such

figures. One possibility is to take the minimum wage rate, if any, times the total number of hours worked in the economy, and assume that compensation of employees

in excess of this estimate represents return to human capita!. Another alternative is to assume that the wage rate in agriculture represents the average compensation of unskilled labor and estimate L as average compensation per employee in agriculture times total employment in the economy. Again compensation of employees in excess

Commodity prices are usually deflated by the manufacturing unit value index. However, since commodities are important inputs in the manufacturing proeess, the manufacturing unit value index is highly affected by the commodity price index. Therefore, to assess the impact of commodity prices on aggregate economic activity, the GDP deflator is the appropriate deflator.

of this is assumed to be return to human capital. In table 1 below the latter alternative

ischosenduetoavailabilty oLdata.8 The countries inc1uded in the table are all

industri

al raw materials importing countries. Apparently the value of ß varies

substantially among countries although it turns out that the value is fairly stable between 0.1 and 0.2 over time in the high-income OECD countries included. Except

for the first two columns, all entries In the table are given as average annual

percentage change over the period.

Table 1

Country ß 1970 ß 1980 avg. 11 avg. 11 avg. -ß11 avg. -ß11 Growth Growth 1970-80 1980-92 1970-80 1980-92 GDP per GDP per (percent (percent (percent (percent capita, capita change) change) change) change) 1970-80 1980-92

Kenya 0.68 0.69 1.3 0.5 -0.9 -0.4 4.4 0.0

Japan 0.30 0.14 -2.1 -8.6 0.6 1.2 3.2 3.4

Korea 0.19 0.28 -1.9 -7.3 0.4 2.1 6.1 7.7

Taiwan 0.17 0.46 -0.7 -7.3 0.1 3.3 7.1 5.9

Denmark 0.12 0.12 -2.1 -6.5 0.3 0.8 1.6 1.8

Ireland 0.17 0.17 -1.6 -6.1 0.3 1.0 3.1 2.9

Before we go on to analyze these results, it is important to note that real local costs of

raw materials depend on the real exchange rate and the co st of local inputs, which to a

large extent mirror relative productivity growth. Therefore, countries which have

experienced dec1ining overall productivity over time, such as Kenya, have seen the

real co st of raw materials rising in spite of the trend depicted in figure 1. Even with

this qualification in mind, table 1 suggests that raw materials prices indeed contribute to different growth rates among countries

8 For Taiwan, compensation of employees by sector is not available, therefore a proxy taking value added in agriculture times the average share of labor in value added for the entire economy is adopted.

Sources for calculations are the following: Commodity prices: World Bank (1994), exchange rates and

Korea and Taiwan represent resource-poor newly industrialized countries who have gained the most fromdecliningcosts of industrial raw materials. First, structural

changes, possibly induced by changes in the relative marginal productivity of capital among sectors due to changes in relative input prices, have induced a shar increase in ßbetween 1970 and 1980, paricularly as far as Taiwan is concerned. Second, the real

cost of industrial raw materials has dec1ined sharly, especially during the L 980s,

inducing growth in per capita income to the tune of on average 2.1 and 3.3 percent during the 1980s for Korea and Taiwan respectively. In other words, if Korea and Taiwan were moving along a steady state growth trajectory during this period, GDP per capita would have grown by 2.1 and 3.3 percent annually in the absence of

technological progress.

OECD countries like Denmark and Ireland have had a stable ß since 1970 and declining real cost of industrial raw materials have generated moderate rates of per capita growth. Finally, Japan has developed from the Korea/aiwan type economy with a relatively high value of ß in 1970 to a typical OECD economy with a relatively low value of ß during the period analyzed. Hence Japan has experienced similar gains as each of the two groups in the respective periods.

Jf the se proxies for ß are anywhere near the actual figures, it is c1ear that commodity

prices have indeed been an important driving force for economic growth. Furthermore the results go some way in explaining how terms of trade are both significant and

important in explaining growth through labor-augmenting cost reductions. Note

finally that the ranking of countries according to actual growth and predicted growth

by the model developed here is similar during the 1980s, when the impact of terms of

trade seems to have be en most profound.

4.2 Foreign direet investment and loeal investment in immobile eapital

In this section empirical evidence for the complementarity between K- and H-capital emphasized in section 3.2 is sought. The model prediets that the net stock of foreign direct investment (FDI) is negatively correlated to the ratio k/, and positively related to real changes in the relative price of imported intermediate goods. Data for the latter variable are not available, and changes in terms of trade (of all merchandise trade) tumed out not to have a significant impact on net FDI. The stock of net FDI per unit of GDP was regressed on the ratio k/, and a number of other variables thought to influence the stocks and flows of FDI. k is taken to be the stock of physical capital per capita owned by domestic citizens, while h is the stock of human capital as

measured by the average number of years of school ing in total population over the age

of 15. Since these two variables are not measured by the same unit of account, indices setting the value in the United States in 1985 = 100 are applied. It turned out that apar from k/ only the share of fuels, metals and minerals in total exports improved the adjusted R square while maintaining an F-value corresponding to a significance

level below 5 percent. The results are shown in table 2 where t-statistics are in

parentheses. 9

Table 2 FDI per unit of GDP

Intercept

k/

^R2 Number of

observations share fuels, metals and

minerals in exports

0.037 -0.0268 0.0476

(2.45) (-2. 17) (1.835)

Sources: k: Penn World Table version 5.6, h: Barm and Lee (1993), share fuels etc.: World Bank (1987), net FD1: UN (1995)

0.145 54

The regression is run on a cross-section of data in 1985, applying OLS.IO The significantly positive coeffcient on the share of fuels, metals and minerals in exports is probably due to the dominance of multinational enterprises in the mineral sector (Dunning 1993). As expected, the coeffcient on k/ is negative and significant. The overall explanatory power of the regression as measured by R2 is relatively low.

However, FDI was heavily regulated in most countries during the period of time of relevance to the regression. Furthermore, FDI is affected by a host of variables related

to investor confidence not incorporated in the mode!. Therefore, a relatively low R2

should be expected.

The theoretical as well as the empirical findings of this paper are related to De Long and Summers (1991 and 1993). They c1aim that accumulation of machinery and equipment, which is probably the most mobile category of capital, is the driving force behind economic growth while they find that the contribution of non-equipment investment to growth is at best minor. In contrast, the model developed here predicts that output is determned by accumulation of H-capital through equation (lb'), albeit

10 1985 is chosen for several reasons: this is he last year for which comprehensive data on human capital is available from Barro and Lee (1993), and it is the base year for constant international price estimates of all the variables in the Penn World tables version 5.6. Data given at current USD from other sources are therefore applied without any adjustments. The 54 countries inc\uded are those for which both data on capital stock are available in the Penn World Tables and data on average number of

years of school ing are available from Barro and Lee (1993). Two outlayers, Zimbabwe and Panama,

were take n out of the sample.

I

there is indeed a one to one relationship between accumulation of K-capital and output

givenbyequation (12). The minorucontribution fromnon-equipment investment in

developing countries found in De Long and Summers (1993) can be explained if large

(public) investments in human capital or infrastructure are accompanied by

restrictions on international capital flows. This could envis age a situation where hi :;

(1-y)/y, which implies that the return to H-capital would go below the international rate of return, and thus below its opportunity cost, while stil improving the rate of

return on K-capita!. 1 i Finally, other studies, all in industrialized OECD countries

report a significantly positive impact of public investment in infrastructure on

productivity in the private sector (Aschauer 1989; Munnell 1990; and Nadiri and

Mamuenas 1994).

5. Summary and conclusions

This paper has analyzed the repercussions of a long-run trend of dec1ining raw

materials prices on commodity-importing countries' productivity and growth, by incorporating raw materials as an essential input in the production function. It is found that dec1ining costs of raw materials "augment" non-accumulated factors in the

same way as technological progress, and induce growth in real output. Further, since

countries differ largelyas far as raw materials iritensity is concerned, dec1ining real

co st of commodities induces international capital flows and differing growth rates among countries even in steady state. In a world of constrained international capital

mobilty, the model predicts that relatively open, resource-intensive economies with a

high investment rate in H-capital wilgrow faste st andattract thejargest inflows of

foreign investment. Recall, however that resource intensity refers to the relative importance of raw materials compared to raw labor, not the use of raw materials as a share of GDP or gross output.

Acknowledgments

This paper has benefited from comments by Kjell Erik Lommerud, Trond Olsen, paricipants at the workshop "Economic Growth and Poverty: The Roles of the

Market, Public Action and the Community", organized by CRED, Namur University, Belgium 6-13 July 1996, and a seminar audience at the University of Bergen.

Referenees:

Ashauer, D.A. (1989), "Is Public Expenditure Productive?", Journal of Monetary

Economics, VoL. 23, pp. 177-200.

Baro, RJ. and J.W. Lee, (1993), "International Comparison of Educational Attainment," NBER Working paper No. 4393.

Baro, R J., N. G. Mankiw and X. Sala-i-Marin (1995): "Capital Mobility in Neoc1assical Models of Economic Growth", American Economic Review, VoL. 85, No. 1, pp. 103-115.

Baro, R J.and X. Sala-i-Marin (1995): "Economic Growth", New York: McGraw-

Hil, Inc.

Baro, RJ. (1996), "Democracy and Growth," Journal of Economic Growth, 1, 1

(March) pp.1-27.

Bruno, M. and J.D. Sachs (1985): "Economics of Worldwide Stagfation", Cambridge, Massachusetts: Harard University Press.

Christini, A., (1995): "Primar Commodity Prices and the OECD Economic

Performance." European Economic Review, VoL. 39, No 1, pp 83-98.

Council for Economic Planning and Development, Republic of China (1994),

"Taiwan Statistical Data Book 1994"

Cuddington, J. T., (1992): "Long Run Trends in 26 Primar Commodity Prices", Journal of Development Economics, VoL. 39, pp 207-227.

De Long, J. B. and L. H. Summers (1991), "Equipment Investment and Economic

Growth", Quarterly Journal of Economics, VoL. 106, No.2, pp 445-502.

De Long, J. B. and L. H. Summers (1993), "How Strongly Do Developing Countries

Benefit from Equipment Investment?", Journal of Monetary Economics, VoL.

32, No. 3, pp. 395-415.

Dunning, J.H., (1993), "Multinational Enterprises and the Global Economy", Wokingham, England: Addison-Wesley Publishing Company.

Grossman, M.G. and E. Helpman (1991), "Innovation and Growth in the Global Economy," Cambridge, Mass.: MIT Press.

IMF: International Financial Statistics, Yearbook 1995.

Munnell, A. H. (1990), "Why Has Productivity Growth Dec1ined? Productivity and Public Investment", New England Economic Review, Janleb, pp. 3-22.

Nadiri, M.I. and T.P. Mamuenas (1994), "The effects of public infrastructure and R&D Capital on the Cost Structure and Performance of U.S Manufacturing Industries", The Review of Economics and Statistics, VoL. 76, No.1, pp. 22-37 NBER, (1995): Penn World Table 5.6

Nordås, H. K. (1996b): "South African manufacturing industries - catching up or falling behind?" Journal of Development Studies, VoL. 32, No. 5, lune, pp.

715-733.

OECD, (1982): "Historical Statistics 1960-1980", Paris: OECD.

Rodriguez-Clare, A., 1996, "The division of labor and economic development",

Journal of Development Economics, VoL. 49, 3-32.

Rodrik, D. (1996): "Coordination failures and government policy: A model with applications to East Asia and Eastern Europe," Journal of International Economics, 40 1-22.

Romer, P.M. (1987) "Growth Based on Increasing Returns Due to Specialization,"

American Economic Review, 77, 2, 56-62.

United Nations, "Yearbook of National Accounts Statistics 1980" and 1989. New York: UN.

United Nations, (1995), "World Investment Report. Transnational Corporations and Competitiveness," New York: United Nations.

World Bank, (1987), "World Development Report 1987," New York: Oxford University Press.

World Bank (1994): Market Outlook for Major Primar Commodities. Report No.

814/94.

World Bank (l 995), "World Data" (database provided on CD-rom)

Recent Working Papers WP 1996: 2

WP 1996: 3

WP 1996:4

WP 1996: 5

WP 1996: 6

WP 1996: 7

WP 1996: 8

WP 1996: 9

WP 1996: 10

WP 1996: 11

WP 1996: 12

WP 1996: 13

WP 1996: 14

WP1997: L

WP 1997: 2

WP 1997: 3

WP 1997: 4

TJOMSLAND, Marit

A discussion of three theoretical approaches to modernity: Understanding

modernity as a globalising phenomenon. Bergen, July 1996,24 pp.

KNUDSEN, Are J. and Kate HALVORSEN

lncome-generating programmesin Pakistan and Malawi: A comparative review.

Bergen, August 1996.

FJELDSTAD, Odd-Helge

Bistand og næringsutvikling. Noen kommentarer til St. meld. nr. 19 (1995-96).

Bergen, august L 996, 17 s.

RAKNER, Lise and Tor SKÂLNES

Political institutions and economic reform: Zambia and South Africa in

comparative context. Bergen, September 1996,44 pp.

RAKNER, Lise

Rational choice and the problem of institutions. A discussion of ratio nal choice

institutionalism and its application by Robert Bates. Bergen, September 1996, 31 pp.

ARIFIN, Bustanul

Economic analysis of land degradation in Indonesia upland. Bergen, September 1996,20 pp.

PAUSEWANG, Siegfried

"Eneweyay" (Lets Discuss). Bergen, September 1996.

ANGELSEN, Arild

Deforestation: Population or market driven? Diferent approaches in modellng agricultural expansion. Bergen, September 1996,30 pp.

ANGELSEN, Arild and Richard Culas

Debt and deforestation: A tenuous link. Bergen, September 1996, 29 pp.

KNUDSEN, Are J.

Deforestation and entrepreneurship in the North West Frontier Province, Pakistan. Bergen, October 1996, 22 pp.

NORDÁS, Hildegunn Kyvik

Trade and growth with static and dynamic economies of seale. Bergen,

November 1966, 36 pp.

LØNNING, Dag Jørund

Dealing with the good and the evil. lntroducing morality as an anthropological concern Bergen, December 1996,27 pp.

FJELDSTAD, Odd-Helge

Tax evasion and corruption in local governments in Tanzania: Alternative economic approaches. Bergen, December 1996,33 pp.

SUMAILA, Rashid Ussif

Strategic dynamic interaction: The case of Barents Sea fisheries. Bergen,

January 1997.

BØE, Turid

Share-tenancy within the household unit. Bergen, February 1997,27 pp.

BØE, Turid

Evolutionary game theory and the battle ofthe sexes. Bergen, March 1997, 16 pp.

BØE, Turid

Teoriskapt virkelighet? Ny-klassisk teori og kvinnelige økonomiske aktører.

Bergen, mars 1997, 17 s.